Filters








99 Hits in 4.6 sec

SenSeI: Sensitive Set Invariance for Enforcing Individual Fairness [article]

Mikhail Yurochkin, Yuekai Sun
2021 arXiv   pre-print
In this paper, we cast fair machine learning as invariant machine learning. We first formulate a version of individual fairness that enforces invariance on certain sensitive sets.  ...  We then design a transport-based regularizer that enforces this version of individual fairness and develop an algorithm to minimize the regularizer efficiently.  ...  ENFORCING INDIVIDUAL FAIRNESS WITH SENSITIVE SET INVARIANCE (SENSEI) 2.1 A TRANSPORT-BASED DEFINITION OF INDIVIDUAL FAIRNESS Let X and Y be the space of inputs and outputs respectively for the supervised  ... 
arXiv:2006.14168v2 fatcat:2exdr4hhifbmnmpfulua2tdw6q

Domain Adaptation meets Individual Fairness. And they get along [article]

Debarghya Mukherjee, Felix Petersen, Mikhail Yurochkin, Yuekai Sun
2022 arXiv   pre-print
methods for domain adaptation to enforce (individual) fairness.  ...  In particular, we show that (i) enforcing suitable notions of individual fairness (IF) can improve the out-of-distribution accuracy of ML models, and that (ii) it is possible to adapt representation alignment  ...  E X∼P [ X − T (X) ] ≤ ǫ. (2.10) This regularizer is similar to the SenSeI regularizer originally proposed and studied by Yurochkin et al. [3] for enforcing individual fairness.  ... 
arXiv:2205.00504v1 fatcat:gnusztq6ubgc3pbzz3qwu5zdim

Individually Fair Gradient Boosting [article]

Alexander Vargo, Fan Zhang, Mikhail Yurochkin, Yuekai Sun
2021 arXiv   pre-print
We consider the task of enforcing individual fairness in gradient boosting.  ...  Gradient boosting is a popular method for machine learning from tabular data, which arise often in applications where algorithmic fairness is a concern.  ...  Unlike other methods for enforcing individually fairness, our approach handles non-smooth ML models such as (boosted) decision trees. 2.  ... 
arXiv:2103.16785v1 fatcat:brvulcg5hzcqjpuioo423ple3u

11. Treatment and Intervention for Children with Developmental Disabilities [chapter]

Junko Teruyama
2020 Child's Play  
On the other hand, the time/space organization within the remedial education setting clearly enforced a certain mode of governmentality.  ...  "It's not fun. " After a moment of silence, Yamada-sensei said, "Alright then. Thank you for telling me.  ... 
doi:10.1525/9780520968844-014 fatcat:kv2zefdap5eudj3m6xylm5sr4i

Treatment and Intervention for Children with Developmental Disabilities [chapter]

Junko Teruyama
2017 Child's Play: Multi-Sensory Histories of Children and Childhood in Japan  
On the other hand, the time/space organization within the remedial education setting clearly enforced a certain mode of governmentality.  ...  "It's not fun. " After a moment of silence, Yamada-sensei said, "Alright then. Thank you for telling me.  ... 
doi:10.1525/luminos.40.l fatcat:nupzh5istbfmvjhaexomy5pumq

iFlipper: Label Flipping for Individual Fairness [article]

Hantian Zhang, Ki Hyun Tae, Jaeyoung Park, Xu Chu, Steven Euijong Whang
2022 arXiv   pre-print
In particular, we show that label flipping is an effective pre-processing technique for improving individual fairness.  ...  Experiments on real datasets show that iFlipper significantly outperforms other pre-processing baselines in terms of individual fairness and accuracy on unseen test sets.  ...  SenSeI [49] enforces invariance on certain sensitive sets and uses a regularization-based approach for jointly optimizing accuracy and fairness, and BuDRO [43] extends the fairness loss used in SenSR  ... 
arXiv:2209.07047v1 fatcat:4bdebjm3vnhj3kh23ixi4rgjwu

Algorithmic Fairness Datasets: the Story so Far [article]

Alessandro Fabris, Stefano Messina, Gianmaria Silvello, Gian Antonio Susto
2022 arXiv   pre-print
We discuss different approaches and levels of attention to these topics, making them tangible, and distill them into a set of best practices for the curation of novel resources.  ...  Secondly, we document and summarize hundreds of available alternatives, annotating their domain and supported fairness tasks, along with additional properties of interest for fairness researchers.  ...  Acknowledgements The authors would like to thank the following researchers and dataset creators for the useful feedback on the data briefs: Alain Barrat, Luc Behaghel, Asia Biega, Marko Bohanec, Chris  ... 
arXiv:2202.01711v3 fatcat:kd546yklwjhvtkrbhtzgbzb2xm

Discriminative Learning of Similarity and Group Equivariant Representations [article]

Shubhendu Trivedi
2022 arXiv   pre-print
However, being able to learn a similarity-sensitive distance function also presupposes access to a rich, discriminative representation for the objects at hand.  ...  In the first part of the thesis, assuming good representations for the data, we present a formulation for metric learning that makes a more direct attempt to optimize for the k-NN accuracy as compared  ...  points to enforce the margin.  ... 
arXiv:1808.10078v2 fatcat:avvdo3t5yvbfpombson5ad4b4m

THE "JŪDŌ SUKEBEI" PHENOMENON: WHEN CROSSING THE LINE MERITS MORE THAN SHIDŌ [MINOR INFRINGEMENT] ― SEXUAL HARASSMENT AND INAPPROPRIATE BEHAVIOR IN JŪDŌ COACHES AND INSTRUCTORS

Carl De Crée
2015 Problems of Psychology in the 21st Century  
Jūdō clubs and organizations overall suffer from a lack of expertise, intellectual insight, ethical objectivity, and solid procedures of fairness for both victim and accused, in this way often failing  ...  The sport of jūdō was intended as an activity "for all".  ...  This should be achieved through each individual striving for the utmost morality.  ... 
doi:10.33225/ppc/15.09.85 fatcat:srzq5jku4zhftbenhhujos2sfu

Algorithmic Fairness Datasets: the Story so Far [article]

Alessandro Fabris, Stefano Messina, Gianmaria Silvello, Gian Antonio Susto
2022
We discuss different approaches and levels of attention to these topics, making them tangible, and distill them into a set of best practices for the curation of novel datasets.  ...  In this work, we survey over two hundred datasets employed in algorithmic fairness research, producing standardized and searchable documentation for each of them, along with in-depth documentation for  ...  Acknowledgements The authors would like to thank the following researchers and dataset creators for the useful feedback on the data briefs: Alain Barrat, Luc  ... 
doi:10.48550/arxiv.2202.01711 fatcat:mav36x3w5namjhurzpevtsmsju

Flexible Scheduling in Middleware for Distributed rate-based real-time applications - Doctoral Dissertation, May 2002

Christopher D. Gill
2002
ARMADA: The ARMADA project [79, 1] defines a set of communication and middleware services that support fault-tolerant and end-to-end guarantees for real-time distributed applications.  ...  The Best-Effort Real-Time (BERT) Scheduler [10] implemented for Scout establishes two classes of service, one with timeliness requirements and one with progress and fairness requirements.  ...  This helps to define and enforce arbitrary partial orderings on tuples based on the available rates, which are essential to policies like FAIR and CB-FAIR.  ... 
doi:10.7936/k7222s4m fatcat:ybualwt7w5gunbqppc7hqd4xti

The Perception of Work in Tokugawa Japan: A Study of Ishida Baigan and Ninomiya Sontoku

Anne Walthall, Eiji Takemura
1999 Journal of Japanese Studies  
Ishida's theories on cultivation invariably had major implications for his views of the Tokugawa social structure.  ...  He argued that both were the product of devoted action to one's occupation, and thus it was neither legitimate nor fair to castigate merchants for their profit-making.  ... 
doi:10.2307/133371 fatcat:nqjn6gij7rcvlhf6yw5dgrhwj4

More Physics Tomorrow

Richard C. Raymond
1949 Physics today  
have to be enforced, or else that it should be eliminated altogether.  ...  The meter is a sensitive, yet rugged instrument, with an expanded, easily-read scale. The maximum light sensitivity is 10 micro lumen. The DENSICHRON is supplied with a blue sensitive light probe.  ... 
doi:10.1063/1.3066461 fatcat:5wzhpz4y3jgmxftqrsbfvzmtb4

The Principles of Deep Learning Theory [article]

Daniel A. Roberts, Sho Yaida, Boris Hanin
2021 arXiv   pre-print
We explain how these effectively-deep networks learn nontrivial representations from training and more broadly analyze the mechanism of representation learning for nonlinear models.  ...  The result highlights that the statistics of the network sensitively depend on the setting of the initialization hyperparameters, with the sensitivity increasing exponentially with depth.  ...  Take a few moments for reflection. . . . . . . . . . Good. You are now a Gaussian sensei. Exhale, and then say as Neo would say, "I know Gaussian integrals."  ... 
arXiv:2106.10165v2 fatcat:l4p4w2gwa5b45lg27wlh2zqncu

Power and pork: a Japanese political life

2008 ChoiceReviews  
The LDP agricultural leaders appealed for sufficient guarantees for a number of sensitive items, and for the establishment of rules to lower tariffs that applied to sensitive items less than those that  ...  Most people were voting for Hosokawa sensei [who was standing for the New Frontier Party].  ... 
doi:10.5860/choice.45-2849 fatcat:6ooqdqpn7vgrjpikfhiyfqllie
« Previous Showing results 1 — 15 out of 99 results